Look, I know what you are here for. You want me to use my powers of industry awareness, combine them with my proximity to VMware, EMC and Cisco and tell you exactly what the acquisition of Whiptail means, not just for VCE but for the industry as a whole, right?
Well, good luck. I counted no less than 5 different NDAs (yes, there are two from Cisco) that prevent me from doing that today. And, the truth is that I’m not “the” CTO of VCE, I’m just one of his stunt doubles. What I mean to say is that my insight into the actual inner workings is highly selectively filtered, and as such I definitely don’t know as much as I want to know, or as much as I could know considering our proximity. I’m willing to accept maybe I know as much as I should, but that’s another story. Sometimes I feel very much like the Genie from Aladdin.
That said, I have lots of opinions, and as I figure a way to tap dance between the legal raindrops of the various NDAs that I am covered under, I’ll do my best to share those. Needless to say, I’m not one of the people that believes this is no big deal. The “Internet of Everything” depends on everything in the data center being a service that can be delivered, and Cisco is, and always has been, a service delivery company. We saw video as a service to be delivered, we saw security as a service to be delivered, we saw compute as a service to be delivered, why would storage be any different? Theoretically. Damn, tap dancing is hard.
While I figure out how to not get fired and still provide some value, let’s talk about the other side of this acquisition. Chad Sakac, Jason Nash and others have speculated on why Cisco acquired Whiptail and not any of the other startups in this space, but no one has really looked at this from the point of view of their existing customers. What happens to them?
Let me tell you a story:
In 2006, I was a huge Compellent fan. I inherited a bunch of it that was being used as target disk for backups at the service provider I worked for, and the amount of functionality that the platform had, even then, was pretty amazing. Storage tiering, thin provisioning, continuous snapshots, there were a lot of things that Compellent got right early on that started to show up in products from EMC and IBM much later.
The thing was, the hardware was the worst kind of shit imaginable. The Series 20 was literally one of the worst designed pieces of white-box trash I ever racked in a data center. One example: it had three power supplies, but needed two of them to run. So how did you cable it up in a data center with redundant feeds? Well, you didn’t. Any power failure on the side with two connections was going to bring the SAN down. Oops. Yes, I found out about this after a power leg failed. See my scars and know I’ve been there.
Around 2008 we started hearing rumors (and by that I mean their management was openly discussing it with customers) that Compellent was looking to be bought, and the option that provided the most value to everyone involved was to get picked up by a storage company that needed the software, but who had their own hardware. After hearing this, I immediately made the strategic decision to stop buying Compellent gear. With no assurance that the hardware I was buying would continue to be supported by whoever bought the company, the risk of being stranded was higher than I was willing to accept. We had a brief (and spectacularly disastrous) flirtation with HP storage before settling on EMC, which remains a strategic partner with Peak 10 to this day.
So my question today, with Cisco being very clear that they are buying Whiptail for the software stack and IP in order to integrate it into the UCS stack, is this: How do you think all of the Whiptail customers are feeling right about now?
In IT, there’s always a risk of product obsolescence. Hell, Google kills off products in use by millions of people without blinking, and EMC refreshes hardware at a sometimes alarming rate. But storage is an interesting beast in that once it’s on the floor and customers have data on it, the expectation is that it’ll work forever. Did you know that EMC will support every model of storage array they have ever sold, no matter how old? It’s one of the costs of doing business in the enterprise storage space. So it’s reasonable to expect that when Whiptail customers bought multiple nodes of data center storage, their expectation was that they’d get a number of years of service out of the platform. Further, they probably expected that there would be a feature growth curve, be it support for new hardware, new software features or other ways to leverage the management and operational accommodations they made when the hardware hit the floor. Training people and changing processes is expensive, and can easily cost as much as the initial capital outlay.
So, if I’m a Whiptail customer today, while I’m happy for the people, I might be a little mad at the stated direction that Cisco is taking. Yesterday, I had a partnership with a company that offered both hardware and software, and today I see that they are being bought as a software layer for a server manufacturer. Maybe Cisco will continue to sell the appliances for a while, heck maybe customer will even continue to buy them, but the Whiptail that people invested in is no longer. That’s not a value judgment on my part (hardware/storage startups out there: when in doubt take the money!), it’s just the truth of the situation. Customers get to be closer to the bleeding edge, customers may get more features for less money, customers get a different kind of relationship with a startup sales team, but the risk is that their exit strategy doesn’t necessarily have to be in anyone’s best interest but their own.
If you look around the industry today, there are a couple companies that are insulated against this risk, just because of the nature of what they do. Nexenta, for example, is, and always has been software only, and unless they get acquired for a specific piece of their stack, customers should have a fairly easy transition. Violin Memory is the other side of that coin, offering essentially a hardware-only solution which should also age with reasonable grace for customers. VMware is another software only offering, and their track record of playing nice with whatever hardware you choose to use has let customers make strategic partnership changes without hurting their operational processes in place upstream.
But there are a lot of software only startup companies that, like Whiptail, are including generic, undifferentiated hardware for different reasons, some of them good ones. Specifically, the folks like Simplivity, Nutanix, Pure, Nimble, Tintri (and the list goes on) are, at their core (IMO), software stacks, and the software is the value they offer, both to customers and investors! I’m not being critical: I understand and agree with many of the reasons for including hardware. Software abstraction is hard; software abstraction on top of a variable hardware stack is extremely difficult. Large enterprises may have a significant existing investment and/or partnership on the hardware side that needs to be respected, but smaller businesses just want things to work out of the box, and including hardware, even low-end hardware, makes that possible. It makes support easier, it makes deployment easier, it boosts top line revenue numbers, all good things to consider when you are a startup looking to attract more investment or execute on an exit strategy.
Please understand, I’m not throwing stones here: the ruthless standardization of hardware is something that provides value to every company, and by extension to their customers. EMC has done it, to fantastic effect. VCE uses standardization as the core of our company. The standardization on x86 processors has been one of the most impactful events in the history of computing. There are very, very difficult challenges and use cases where standardization is to key to minimize the effort needed to provide value to customers. I get it. I’m a drinker of the Kool-Aid.
But there’s risk, if you are a customer, and the Whiptail deal may show how that risk can manifest. Would it be enough to make me choose not to purchase one of those platforms? Probably not, depending on where they are in the funding process and how many rounds in they were. But it would certainly be something I’d ask about, and certainly something that would get factored in. Because tomorrow that all-in-one hardware platform may be a software storage stack that Lenovo or IBM use to create scale out blade systems. And then where are you?
I know, this wasn’t the Whiptail truth you were looking for. But having come out of the service provider space where I felt the pain of startups being acquired, I think it’s an important truth to remember. Those of us on the vendor side of the table revel in and celebrate when a company has an equity event, but to customers it’s not always a great thing.
Thoughts? Comments? Have you been a customer of a company that was acquired? How did it affect your relationship with the vendor? Courteous discussion welcome below, or on Twitter!